Director of Engineering Physics and Associate Professor, Department of Physics and Astronomy
University of British Columbia, Vancouver
Name: Andre Marziali
Title: Director of Engineering Physics and Associate Professor, Department of Physics and Astronomy, University of British Columbia, Vancouver, since 2005
Experience and Education:
- Assistant professor, University of Vancouver, 1998-2005
- Lead research engineer, Ron Davis’ group, Stanford University, 1994-1998
- PhD in Physics, Stanford University, 1994
- BASc in Engineering Physics, University of British Columbia, 1989
Earlier this month, Andre Marziali won a three-year grant under the National Human Genome Research Institute’s Advanced Sequencing Technology program to develop a nanopore array force spectroscopy chip for rapid clinical genotyping (see In Sequence 8/7/2007).
His group, working under a 2004 feasibility grant obtained through the same program, already showed that its nanopore-based method can resolve single nucleotide differences between target DNA molecules. In Sequence caught up with Marziali last week to find out what he has left to do.
How does your genotyping method using nanopore-based force spectroscopy work?
Rather than reading DNA sequences that pass through the pore, we use a nanopore to confine single-stranded DNA. We take advantage of the fact that we can tell whether a nanopore is occupied with DNA or not. That signal is very strong; it’s very easy to tell whether there is a DNA molecule in the pore or not. What’s hard to do right now is to tell what the sequence of the DNA inside the pore is. We have taken the easier approach of simply detecting whether there is DNA in the pore or not.
The way we detect the sequence is by putting a certain sequence of single-stranded DNA through the pore, so it protrudes to the other side of the pore. We then bind an analyte that may or may not match that sequence. We are really looking for a single-base mutation in known sequence; we are not sequencing de novo. Once the analyte binds, it forms a duplex. By using electric fields, we then try to pull that duplex through the pore. The pore is too small for the duplex to fit, so for the probe molecule that we inserted to escape the pore, the duplex has to be ripped apart. And the amount of force required to rip the duplex apart tells us whether the match between the analyte and the probe is a perfect match, or whether there is a single-base mismatch.
We have published two papers showing the signal-to-noise [ratio] with which we can detect single-base mismatches. It allows us, in a single, multipore measurement, to detect even single-base mutations in a short sequence.
The goal is to eventually develop a synthetic nanopore array that would be part of a small microfluidic chip that would allow you to analyze DNA for, say, dozens to maybe a hundred different mutations. In a single, rapid measurement [you would] be able to tell whether those 100 loci [matched] a specific allele that might be indicative of a disease or a response to a certain drug, or something of clinical relevance.
So you are now switching your method from organic to synthetic pores?
With the organic alpha-hemolysin pore, the signal-to-noise ratio is fantastic, and we have shown in that pore that you can achieve single-nucleotide resolution. Unfortunately, that pore is hard to form, and it forms in a lipid bilayer, which is not very stable.
So what we have been doing is developing synthetic pores, and arrays of synthetic pores. These are essentially nanopores fabricated by lithography and electron beam methods, very similar to the methods used to fabricate electronic circuits.
Part of the challenge that everyone faces who works with synthetic nanopores is to make nanometer-scale pores that have properties similar to the organic pores. We have made great advances in that we have brought the noise levels of synthetic nanopores to within a factor of two of that of organic pores.
The organic pore has the property that its size is just big enough to let single-stranded DNA into it; it’s too small for double-stranded DNA. Controlling the synthetic pores at that level will be difficult. The pores we have are big enough for double-stranded DNA, which will go through this. Now, we have to come up with schemes for attaching proteins and multiple probes, so we can use the synthetic pore like an organic pore. This will involve methods like a sandwich assay where multiple probes bind to the analyte of interest at multiple points. The goal is eventually to detect mutations in samples that have not been PCR-amplified.
What would be the main advantages over conventional genotyping methods?
It would be the advantage of simplicity and speed. Clinical genotyping instruments exist but they can genotype only one mutation in under an hour. Genotyping a number of loci requires technologies that have a longer turnaround time and often require the sample to be shipped outside the clinic. These existing methods are not amenable to the rapid turnaround required in a clinical setting.
Our collaborators are from St. Paul’s Hospital here in Vancouver. They are looking at trying to use genetic markers to guide the treatment of diseases like sepsis. There are conditions where the clinical intervention has to be very rapid — they need genotyping information on the order of minutes.
We are looking to be able to do 100 markers in less than half an hour, and at the same time keep it inexpensive. We would hopefully be able to bypass PCR amplification, and genotype directly from blood samples.
How will your work benefit other groups that are working on nanopore-based sequencing methods?
We just published a paper on noise in synthetic pores and our approaches to reducing that noise in Nanotechnology. We did a careful analysis of where the noise sources were in synthetic pores and figured out how to reduce at least some of the noise sources. I think we actually have the lowest-noise pores of anyone out there right now. That in itself should be of great interest to people using synthetic pores for DNA sequencing because they obviously will also need to reduce their noise.
Other studies we are doing that I think will be of interest to them is the kinetics of DNA translocation through the pores. A lot of these schemes will depend heavily on understanding how fast the DNA goes through the pore, and not only that, but from one molecule to the next, what is the difference in the rates at which the molecules might go through. We found in the organic pores that there is a tremendous difference from one molecule to the next. There are molecules that go through in a millisecond, and then other molecules will take ten seconds to go through. We have actually studied that in detail and shown that some of the assumptions that were been made in terms of the distribution of times for the molecule passage were incorrect. We have a far better understanding what the distributions would be now, which will help guide people in development of sequencing detection schemes.
And then in general, the fabrication methods and fabrication of arrays of pores [will be useful to other groups.] Obviously, we share all of these things with the grantees at the NHGRI meetings.
One of the reasons why we are pursuing the direction that we are pursuing, which is not direct sequencing, is that we think, technically, it’s much easier to achieve in the short term. This is not a 10-year horizon, it’s more like a two-year horizon. You never know how things go, but it’s certainly technically easier than sequencing.
What are your plans for commercializing your genotyping method?
We haven’t started on that yet. Only in the sense that we would like to have a proof-of-concept prototype for at least one element of a synthetic chip before we start that process. But the idea is, at that point, to seek a commercial partner that would want to do the fabrication of the entire chip. Our capabilities in fabrication will be limited to what we can do in a university clean room. If this is going to be produced as a consumable for the clinical market, you would have to use larger fabrication facilities. This is where an Agilent or an Intel or someone might be interested in hopefully pursuing this.
Do any of your other projects – small volume liquid handling, biomolecule concentration – have potential applications for DNA sequencing as well?
Yes, the biomolecule concentration, specifically DNA concentration, has applications to sequencing as a front end, as a sample-preparation system. We have a grant application into NHGRI to use that technology for selecting specific sequences within genomic DNA for preparation of libraries to then feed into, say, a Solexa machine or something like that. But it’s not a sequencing technology by any means.
Are you planning to develop this commercially?
We already incorporated a company to commercialize that technology, Boreal Genomics. We are pursuing the development of an instrument as a bench-top DNA purification instrument for the life science research market in general, but also partnering with companies that are interested in using our technology as a front end to a DNA sequencing detection system or things along those lines. That one is actually quite far along; we have commercial partners already and have done a number of proof-of-concept [studies]. We have not published a lot on it, unfortunately. But we are working on a couple of publications now, which hopefully will come out soon.